Intelligent inspection method and system based on artificial intelligence robot

Abstract

The invention relates to the technical field of robot inspection, and particularly discloses an intelligent inspection method and system based on an artificial intelligence robot, and the method comprises the following steps: S1, carrying out the clustering of inspection point locations, obtaining point location clusters, sorting the point location clusters, and taking the first point location cluster in the sorting as a to-be-determined cluster; s2, connecting two adjacent inspection point positions in the to-be-determined cluster through a straight line to obtain an initial route, and taking the initial route with the shortest distance as a target route; determining a target cluster according to the length of the target route and the maximum endurance mileage of the artificial intelligence robot; and S3, removing the inspection point locations in the target clusters, repeating the steps, determining all the target clusters, and determining an inspection plan according to the target clusters. According to the invention, high inspection cost can be avoided while the inspection efficiency is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of robot patrol inspection, and particularly relates to an intelligent patrol inspection method and system based on an artificial intelligence robot. Background Art

[0002] An artificial intelligence robot is an intelligent system that combines artificial intelligence (AI) technology and robot hardware, aiming to enable the robot to have the capabilities of autonomous learning, environment perception, decision-making, and task execution. It obtains external information through sensors, analyzes data using AI algorithms, and makes corresponding decisions or actions. Artificial intelligence robots are widely used in fields such as intelligent patrol inspection.

[0003] In large-scale patrol inspection tasks, the battery life of artificial intelligence robots is a very important issue. Most artificial intelligence robots rely on battery power supply, but the current energy density of battery technology is limited, resulting in limited battery life mileage of artificial intelligence robots, and the need to frequently replace batteries or charge the artificial intelligence robots, which reduces the efficiency of patrol inspection.

[0004] The above problems can be solved by increasing the number of artificial intelligence robots. However, generally, how many artificial intelligence robots to increase is determined based on the experience of managers. This method is not scientific enough. If the number of artificial intelligence robots is small, the patrol inspection efficiency will still decline. If the number of artificial intelligence robots is large, the patrol inspection cost will increase. Therefore, how to balance the patrol inspection efficiency and cost has become an urgent problem to be solved. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent patrol inspection method and system based on an artificial intelligence robot to solve the above technical problems.

[0006] The purpose of the present invention can be achieved by the following technical solutions: An intelligent patrol inspection method based on an artificial intelligence robot includes the following steps: S1: Determine the patrol inspection points in the patrol inspection area, cluster the patrol inspection points to obtain point clusters, sort the point clusters, and the more the number of patrol inspection points in the point cluster, the higher the ranking of the point cluster. The point cluster ranked first in the sorting is used as the pending cluster; S2: Starting from the patrol inspection point A in the pending cluster, connect the adjacent two patrol inspection points in the pending cluster by a straight line to obtain an initial route. The initial route contains all the patrol inspection points in the pending cluster and each patrol inspection point appears only once in the initial route. The initial route with the shortest distance is used as the target route; Determine the target cluster according to the length L of the target route and the maximum battery life mileage Lys of the artificial intelligence robot. The target cluster includes the patrol inspection points for the artificial intelligence robot to perform patrol inspection; S3: Remove the inspection points in the target cluster obtained, repeat the steps S1 - S2 above to determine all target clusters, and determine an inspection plan based on the target clusters. The inspection plan includes the number of AI robots and the inspection path. When the length L of the target route within the target cluster is < 0.5Lys, perform the following steps: Determine the target cluster C where the length L of the target route is < 0.5Lys, determine the distances between the center of the target cluster C and the centers of the other target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and the target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB within the target cluster CCdy, and take all the monitoring points from the endpoint of the target route MB to the reference point K as one target cluster.

[0007] As a further aspect of the present invention: In the step S3, the reference point K satisfies the following constraints: ; where ε is a preset correction coefficient, η fir and η sec are respectively a preset first coefficient and a second coefficient, and the first coefficient η fir < 0, the second coefficient η sec > 0, F and G represent the two endpoints of the target route MB, B F，K represents the length of the target route BFK, the target route BFK represents the target route corresponding to the initial cluster FK, the initial cluster FK represents the cluster of points formed by the inspection points between the endpoint F and the reference point K, S F，K represents the number of inspection points in the initial cluster FK, and S tot represents the total number of inspection points on the target route MB.

[0008] As a further aspect of the present invention: In the process of determining the reference point K, the following steps are further included: When the points satisfying the constraint conditions on the target route MB are not unique, take the points satisfying the constraint conditions as the points to be determined, and determine the undetermined values of the points to be determined , and take the point to be determined corresponding to the undetermined value closest to 0 as the reference point K; When there are no points satisfying the constraint conditions on the target route MB, send a warning message to remind the staff.

[0009] As a further aspect of the present invention: In the step S2, the process of determining the target cluster according to the length L of the target route and the maximum cruising range Lys of the AI robot specifically includes: When the length of the target route L>0.9Lys, perform the following steps: Determine the distance between the inspection point in the pending cluster and the center of the remaining point clusters, and determine the shortest distance Dmin nb , remove the shortest distance Dmin from the pending cluster nb Corresponding inspection points; Re-determine the target route in the pending cluster and repeat the above steps until the length L of the new target route in the pending cluster is ≤ 0.9Lys after the inspection points in the pending cluster are removed, and the pending cluster at this time is used as the target cluster; When the length of the target route L ≤ 0.8Lys, perform the following steps: The inspection points that do not belong to the pending cluster are regarded as vacant points, the distance between the vacant points and the center of the pending cluster is determined, and the shortest distance Dmin is determined. ky , the shortest distance Dmin ky The corresponding vacant points are merged into the pending cluster; The target route in the pending cluster is re-determined and the above steps are repeated until the length L of the new target route in the pending cluster after the vacant points are incorporated is greater than 0.8Lys, and the pending cluster at this time is used as the target cluster.

[0010] As a further solution of the present invention: when the inspection points in the pending cluster are removed for the ath time, the length of the target route is 0.9Lys<L a ≤0.98Lys, and the length of the target route after the next adjacent inspection point in the pending cluster is removed is L a+1 When ≤0.8Lys, the undetermined cluster after removing the inspection points for the ath time is taken as the target cluster.

[0011] As a further solution of the present invention: when the bth vacant point is incorporated, the length L of the target route is b <0.8Lys, and the length of the target route after the next adjacent vacant point is L b+1 When >0.98Lys, the undetermined cluster after the b-th incorporation of the vacant points is taken as the target cluster.

[0012] As a further solution of the present invention: in the step S3, the process of determining the inspection plan according to the target cluster specifically includes: The number of target clusters is used as the number of AI robots, and one AI robot inspects the inspection points within one target cluster; Determine a target route in the target, and use the target route as an inspection path for the artificial intelligence robot during inspection.

[0013] An intelligent inspection system based on an artificial intelligence robot, comprising: Initial module: Determine the inspection points within a preset inspection area, cluster the inspection points to obtain point clusters, sort the point clusters according to a preset rule, and take the point cluster at the first place in the sorting as the pending cluster; Analysis module: Starting from the inspection point A in the pending cluster, connect two adjacent inspection points in the pending cluster by a straight line to obtain an initial route. The initial route contains all the inspection points in the pending cluster and each inspection point appears only once in the initial route. Take the initial route with the shortest distance as the target route; Determine the target cluster according to the length L of the target route and the maximum endurance mileage Lys of the artificial intelligence robot. The target cluster includes the inspection points for the artificial intelligence robot to conduct inspections; Generation module: Remove the inspection points in the target cluster, repeat the steps in the initial module and the analysis module, determine all the target clusters, and determine the inspection plan according to the target clusters. The inspection plan includes the number of artificial intelligence robots and the inspection paths; When the length L of the target route within the target cluster < 0.5Lys, perform the following steps: Determine the target cluster C with the length L of the target route < 0.5Lys, determine the distance between the center of the target cluster C and the centers of the other target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and the target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB in the target cluster CCdy, and take all the monitoring points from the endpoint of the target route MB to the reference point K as a target cluster.

[0014] Advantages of the present invention: In the present invention, first, the inspection points within the preset inspection area are clustered. An artificial intelligence robot only needs to be responsible for the inspection points within one cluster, avoiding having an artificial intelligence robot responsible for inspection points that are far apart, reducing the situation where the battery of the artificial intelligence robot cannot support it to reach the next inspection point. The clustering method can be selected according to the actual situation, such as K-means clustering, etc., and will not be elaborated here. Considering that the number and location distribution of the inspection points within each cluster after clustering may be different, there may still be a situation where an artificial intelligence robot cannot complete the inspection task within the point cluster (that is, all inspection points are inspected once) without replacing the battery and / or charging the battery. Therefore, the target route within the point cluster is determined, that is, the shortest route to complete the inspection task, and the inspection points within the point cluster are increased or decreased according to the length of the target route and the maximum battery life Lys of the artificial intelligence robot, so as to avoid the situation where the battery power of the artificial intelligence robot cannot support the completion of the inspection task and / or the number of points the artificial intelligence robot is responsible for is too small, resulting in an increase in the number of artificial intelligence robots (for example, the inspection points within a certain point cluster are incorporated into other point clusters, and the artificial intelligence robot in the incorporated point cluster can also complete the inspection task without replacing the battery and / or charging the battery. In this case, the number of point clusters is reduced, thereby reducing the number of artificial intelligence robots required); finally, all target clusters are determined, and the inspection plan is determined according to the target clusters; it should be noted that after all target clusters are determined, there may be a situation where the target route within some target clusters is relatively short (that is, the length L of the target route < 0.5Lys). In this case, the artificial intelligence robot within the target cluster may quickly complete the inspection task, while the other target clusters will take more time to complete the inspection task because they do not belong to the above situation. Therefore, it is selected to merge target cluster C and target cluster Cdy into target cluster CCdy, and re-divide target cluster CCdy into two target clusters according to the target route within target cluster CCdy, thereby reducing the average time for the artificial intelligence robots in the original target cluster C and the original target cluster Cdy to complete the inspection task. The present invention can avoid excessive inspection costs while ensuring inspection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 is a flowchart of an intelligent inspection method based on an artificial intelligence robot of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] Please refer to Figure 1 As shown, the present invention is an intelligent inspection method based on an artificial intelligence robot, including the following steps: S1: Determine the inspection points within the inspection area, cluster the inspection points to obtain point clusters, sort the point clusters, and the more the number of inspection points in a point cluster, the higher the ranking of the point cluster. Take the point cluster ranked first in the sorting as the to-be-determined cluster; S2: Take the inspection point A in the to-be-determined cluster as the starting point, and connect two adjacent inspection points in the to-be-determined cluster by a straight line to obtain an initial route. The initial route contains all the inspection points in the to-be-determined cluster and each inspection point appears only once in the initial route. Take the initial route with the shortest distance as the target route; Determine the target cluster according to the length L of the target route and the maximum endurance mileage Lys of the artificial intelligence robot. The target cluster includes the inspection points for the artificial intelligence robot to perform inspections; S3: Remove the inspection points in the target cluster, repeat the steps S1 - S2, determine all the target clusters, and determine the inspection plan according to the target clusters. The inspection plan includes the number of artificial intelligence robots and the inspection paths; When the length L of the target route within the target cluster < 0.5Lys, perform the following steps: Determine the target cluster C with the length L of the target route < 0.5Lys, determine the distances between the center of the target cluster C and the centers of the other target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and the target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB in the target cluster CCdy, and take all the monitoring points from the endpoint of the target route MB to the reference point K as a target cluster.

[0019] It should be noted that first, the inspection points within the preset inspection area are clustered. An artificial intelligence robot only needs to be responsible for the inspection points within one cluster, avoiding having an artificial intelligence robot responsible for inspection points that are far apart, and reducing the situation where the battery of the artificial intelligence robot cannot support it to reach the next inspection point. The clustering method can be selected according to the actual situation, such as K-means clustering, etc., which will not be elaborated here. Considering that the number and location distribution of the inspection points within each cluster after clustering may be different, there may still be a situation where an artificial intelligence robot cannot complete the inspection task within the point cluster (i.e., all inspection points are inspected once) without replacing the battery and / or charging the battery. Therefore, the target route within the point cluster, that is, the shortest route to complete the inspection task, is determined, and the inspection points within the point cluster are increased or decreased according to the length of the target route and the maximum cruising range Lys of the artificial intelligence robot, so as to avoid the situation where the battery power of the artificial intelligence robot cannot support the completion of the inspection task and / or the number of points responsible for the artificial intelligence robot is too small, resulting in an increase in the number of artificial intelligence robots (for example, the inspection points within a certain point cluster are incorporated into other point clusters, and the artificial intelligence robot in the incorporated point cluster can also complete the inspection task without replacing the battery and / or charging the battery. In this case, the number of point clusters is reduced, thereby reducing the number of artificial intelligence robots required); finally, all target clusters are determined, and the inspection plan is determined according to the target clusters; it should be noted that after all target clusters are determined, there may be a situation where the target routes within some target clusters are relatively short (i.e., the length L of the target route < 0.5Lys). In this case, the artificial intelligence robot within the target cluster may quickly complete the inspection task, while the other target clusters will take more time to complete the inspection task because they do not belong to the above situation. Therefore, it is selected to merge target cluster C and target cluster Cdy into target cluster CCdy, and the target cluster CCdy is re-divided into two target clusters according to the target route within the target cluster CCdy, so as to reduce the average time for the artificial intelligence robots in the original target cluster C and the original target cluster Cdy to complete the inspection task; in this solution, the acquisition method of all target routes can be analogized to the acquisition method of the target route in the to-be-determined cluster, which will not be elaborated here.

[0020] In another preferred embodiment of the present invention, in step S3, the reference point K satisfies the following constraints: ; where ε is a preset correction coefficient, η fir and η sec are respectively a preset first coefficient and a second coefficient, and the first coefficient η fir < 0, the second coefficient η sec > 0, F and G represent the two endpoints of the target route MB, B F，KDenote the length of the target route BFK. The target route BFK represents the target route corresponding to the initial cluster FK, and the initial cluster FK represents a cluster of inspection points composed of the inspection points between the end point F and the reference point K, S F，K Denote the number of inspection points in the initial cluster FK, S tot Denote the total number of inspection points on the target route MB.

[0021] It should be noted that it takes a certain amount of time for the AI robot to inspect an inspection point, and it also takes a certain amount of time to move from one inspection point to another. Therefore, to reduce the average time for the AI robot to complete the inspection task, if the proportion of the number of inspection points between the end point F and a certain point is relatively high, then the length of the route between the end point and this point should be shorter, so that some inspection points can reach the other part of the target route MB (i.e., between the reference point and the other end point), and finally find a point that meets the above constraints as the reference point, B F，K ≤0.95Lys and B G，K ≤0.95Lys is to prevent the reference point from being too close to a certain end point, which may lead to too long a distance between the other end point and the reference point, and further lead to the inability of the AI robot in the target cluster to complete the corresponding inspection task.

[0022] In another preferred embodiment of the present invention, in the process of determining the reference point K, the following steps are further included: When the points satisfying the constraint conditions on the target route MB are not unique, use the points satisfying the constraint conditions as the points to be determined, and determine the undetermined values of the points to be determined , and use the point to be determined corresponding to the undetermined value closest to 0 as the reference point K; When there are no points satisfying the constraint conditions on the target route MB, send a warning message to remind the staff.

[0023] It can be understood that using the point to be determined corresponding to the undetermined value closest to 0 as the reference point can ensure the shortest average time for the AI robot to complete the inspection task.

[0024] In another preferred embodiment of the present invention, in step S2, the process of determining the target cluster according to the length L of the target route and the maximum cruising range Lys of the AI robot specifically includes: When the length L of the target route > 0.9Lys, perform the following steps: Determine the distance between the inspection points in the undetermined cluster and the centers of the other point clusters, and determine the shortest distance Dmin nb , and remove the inspection points corresponding to the shortest distance Dmin nb from the undetermined cluster; Re-determine the target route in the pending cluster and repeat the above steps until the length L of the new target route in the pending cluster is ≤ 0.9Lys after the inspection points in the pending cluster are removed, and the pending cluster at this time is used as the target cluster; When the length of the target route L ≤ 0.8Lys, perform the following steps: The inspection points that do not belong to the pending cluster are regarded as vacant points, the distance between the vacant points and the center of the pending cluster is determined, and the shortest distance Dmin is determined. ky , the shortest distance Dmin ky The corresponding vacant points are merged into the pending cluster; The target route in the pending cluster is re-determined and the above steps are repeated until the length L of the new target route in the pending cluster after the vacant points are incorporated is greater than 0.8Lys, and the pending cluster at this time is used as the target cluster.

[0025] It should be noted that, considering the different numbers and location distributions of inspection points in each cluster after clustering, there may be an artificial intelligence robot that is still unable to complete the inspection task in the point cluster without replacing the battery and / or charging the battery (i.e., all inspection points are inspected once). Therefore, the target route within the point cluster is determined, that is, the shortest route to complete the inspection task, and the inspection points in the point cluster are increased or decreased according to the length of the target route and the maximum cruising range Lys of the artificial intelligence robot, so as to avoid the situation where the power of the artificial intelligence robot cannot support the completion of the inspection task and / or the artificial intelligence robot is responsible for too few points, resulting in an increase in the number of artificial intelligence robots (such as the inspection points in a point cluster are merged into other point clusters, and the artificial intelligence robots in the merged point clusters can also complete the inspection task without replacing the battery and / or charging the battery. In this case, the number of point clusters is reduced, thereby reducing the number of artificial intelligence robots required, thereby reducing the cost of inspection).

[0026] In another preferred embodiment of the present invention, when the inspection points in the pending cluster are removed for the ath time, the length of the target route is 0.9Lys<L a ≤0.98Lys, and the length of the target route after the next adjacent inspection point in the pending cluster is removed is L a+1 When ≤0.8Lys, the undetermined cluster after removing the inspection points for the ath time is taken as the target cluster.

[0027] It should be noted that after removing the inspection points in the pending cluster for the ath time, the length of the target route is 0.9Lys<L a≤0.98Lys, which means that after removing the inspection points, the length of the target route remains within a relatively safe range. Because even considering the uncertain factors that may exist in actual operation (such as complex terrain, changes in the movement efficiency of the robot, etc.), such a battery life ratio is sufficient to ensure that the robot completes the inspection task. However, if after removing the inspection points in the next adjacent cluster to be determined, the length of the target route L a+1 ≤0.8Lys, which means that after further removing the inspection points, the length of the target route will drop to less than 80% of the maximum battery life of the robot. In this case, the robot may have a relatively large amount of remaining power after completing the inspection task, which is a waste of resources to a certain extent; therefore, in order to avoid unnecessary resource waste and ensure that the robot has enough power to complete the inspection task, when the above situation occurs, the cluster to be determined after the a-th removal of the inspection points is selected as the target cluster.

[0028] In another preferred embodiment of the present invention, when the length L of the target route after the b-th incorporation of the vacant points b <0.8Lys, and the length L of the target route after the next adjacent incorporation of the vacant points b+1 >0.98Lys, then the cluster to be determined after the b-th incorporation of the vacant points is used as the target cluster.

[0029] In another preferred embodiment of the present invention, in the step S3, the process of determining the inspection plan according to the target cluster specifically includes: Taking the number of target clusters as the number of artificial intelligence robots, and one artificial intelligence robot inspects the inspection points within one target cluster; Determining the target route in the target, and using the target route as the inspection path when the artificial intelligence robot conducts the inspection.

[0030] It is worth noting that by ensuring that each robot is only responsible for one target cluster, task overlap and resource waste between robots can be avoided. This one-to-one allocation method can maximize the working efficiency of each robot, reduce unnecessary energy consumption and equipment wear; directly taking the number of target clusters as the number of artificial intelligence robots can simplify the management and scheduling process. Each robot is responsible for a specific target cluster, avoiding complex allocation and coordination problems, and improving the convenience and efficiency of management; the target route is the shortest initial route, and using the target route as the inspection path when the artificial intelligence robot conducts the inspection can reduce the time and power consumed by the artificial intelligence robot during the inspection process.

[0031] An intelligent inspection system based on an artificial intelligence robot, comprising: Initial module: Determine the inspection points within the preset inspection area, cluster the inspection points to obtain point clusters, sort the point clusters according to preset rules, and take the point cluster at the first place in the sorting as the pending cluster; Analysis module: Starting from the inspection point A in the pending cluster, connect two adjacent inspection points in the pending cluster by a straight line to obtain an initial route. The initial route contains all the inspection points in the pending cluster and each inspection point appears only once in the initial route. Take the initial route with the shortest distance as the target route; Determine the target cluster according to the length L of the target route and the maximum battery life Lys of the artificial intelligence robot. The target cluster includes the inspection points for the artificial intelligence robot to conduct inspections; Generation module: Remove the inspection points in the target cluster, repeat the steps in the initial module and the analysis module, determine all the target clusters, and determine the inspection plan according to the target clusters. The inspection plan includes the number of artificial intelligence robots and the inspection paths; When the length L of the target route within the target cluster < 0.5Lys, perform the following steps: Determine the target cluster C with the length L of the target route < 0.5Lys, determine the distances between the center of the target cluster C and the centers of the other target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and the target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB in the target cluster CCdy, and take all the monitoring points from the endpoint of the target route MB to the reference point K as a target cluster.

[0032] The above has described an embodiment of the present invention in detail, but the content is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the present invention application should still fall within the scope covered by the patent of the present invention.

Claims

1. An intelligent inspection method based on an artificial intelligence robot, characterized in that, It includes the following steps: S1: Determine the inspection points within the inspection area, cluster the inspection points to obtain point clusters, sort the point clusters. The more inspection points there are within a point cluster, the higher the ranking of the point cluster. Take the point cluster ranked first in the sorting as the to-be-determined cluster; S2: Starting from inspection point A in the to-be-determined cluster, connect adjacent two inspection points in the to-be-determined cluster by a straight line to obtain an initial route. The initial route contains all the inspection points in the to-be-determined cluster and each inspection point appears only once in the initial route. Take the initial route with the shortest distance as the target route; Determine the target cluster according to the length L of the target route and the maximum endurance mileage Lys of the AI robot. The target cluster includes the inspection points for the AI robot to conduct inspections; S3: Remove the inspection points in the target cluster, repeat steps S1 - S2, determine all the target clusters, and determine the inspection plan according to the target clusters. The inspection plan includes the number of AI robots and the inspection path; When the length L of the target route within the target cluster < 0.5Lys, execute the following steps: Determine the target cluster C with the length L of the target route < 0.5Lys, determine the distance between the center of the target cluster C and the centers of the remaining target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB in the target cluster CCdy, and take all the monitoring points between the endpoints of the target route MB and the reference point K as one target cluster.

2. The intelligent inspection method based on an artificial intelligence robot according to claim 1, wherein In step S3, the reference point K satisfies the following constraints: ； where ε is a preset correction coefficient, η fir and η sec are respectively a preset first coefficient and a second coefficient, and the first coefficient η fir < 0, and the second coefficient η sec > 0, F and G represent two endpoints of the target route MB, B F，K represents the length of the target route BFK, and the target route BFK represents the target route corresponding to the initial cluster FK. The initial cluster FK represents a point cluster composed of inspection points between the endpoint F and the reference point K. S F，K represents the number of inspection points in the initial cluster FK, and S tot represents the total number of inspection points on the target route MB.

3. The intelligent inspection method based on an artificial intelligence robot according to claim 2, wherein, During the process of determining the reference point K, the following steps are also included: When the points satisfying the constraint conditions on the target route MB are not unique, the points satisfying the constraint conditions are used as the points to be determined, and the undetermined values of the points to be determined are determined. , and the point to be determined corresponding to the undetermined value closest to 0 is used as the reference point K; When there is no point on the target route MB that satisfies the constraint conditions, send a warning message to remind the staff.

4. An intelligent inspection method based on an artificial intelligence robot according to claim 1, characterized in that, In step S2, the process of determining the target cluster according to the length L of the target route and the maximum endurance mileage Lys of the AI robot specifically includes: When the length L of the target route > 0.9Lys, execute the following steps: Determine the distance between the inspection points in the to-be-determined cluster and the center of the remaining point clusters, and determine the shortest distance Dmin among them. nb , Remove the inspection points corresponding to the shortest distance Dmin from the to-be-determined cluster. nb ; Redetermine the target route in the to-be-determined cluster and repeat the above steps until the length L of the new target route in the to-be-determined cluster after removing the inspection points in the to-be-determined cluster for a certain time ≤ 0.9Lys. Take the to-be-determined cluster at this time as the target cluster; When the length L of the target route ≤ 0.8Lys, execute the following steps: Take the inspection points that do not belong to the to-be-determined clusters as free points, determine the distances between the free points and the centers of the to-be-determined clusters, and determine the shortest distance Dmin among them. ky , and take the shortest distance Dmin ky corresponding free points and incorporate them into the to-be-determined clusters. Redetermine the target route in the to-be-determined cluster and repeat the above steps until the length L of the new target route in the to-be-determined cluster after incorporating the vacant points for a certain time > 0.8Lys. Take the to-be-determined cluster at this time as the target cluster.

5. The intelligent inspection method based on an artificial intelligence robot according to claim 4, wherein When the length of the target route Lys after removing the inspection points in the to-be-determined cluster for the a-th time satisfies 0.9Lys < L a ≤ 0.98Lys, and the length of the target route L a+1 ≤ 0.8Lys after removing the inspection points in the to-be-determined cluster for the next adjacent time, then the to-be-determined cluster after removing the inspection points for the a-th time is taken as the target cluster.

6. The intelligent inspection method based on an artificial intelligence robot according to claim 4, wherein, The length L of the target route after the b-th incorporation into the vacant point b <0.8Lys, and the length L of the target route after the next adjacent incorporation into the vacant point b+1 >0.98Lys, then the pending cluster after the b-th incorporation into the vacant point is taken as the target cluster.

7. An intelligent inspection method based on an artificial intelligence robot according to claim 1, characterized in that, In step S3, the process of determining the inspection plan according to the target clusters specifically includes: Take the number of target clusters as the number of AI robots, and one AI robot conducts inspections on the inspection points within one target cluster; Determine the target route in the target, and take the target route as the inspection path when the AI robot conducts inspections.

8. An intelligent inspection system based on an artificial intelligence robot, characterized in that, It includes: Initial module: Determine the inspection points within the preset inspection area, cluster the inspection points to obtain point clusters, sort the point clusters according to the preset rules, and take the point cluster ranked first in the sorting as the to-be-determined cluster; Analysis module: Starting from the inspection point A in the to-be-determined cluster, an initial route is obtained by connecting two adjacent inspection points in the to-be-determined cluster with a straight line. The initial route contains all the inspection points in the to-be-determined cluster and each inspection point appears only once in the initial route. The initial route with the shortest distance is used as the target route; Determine the target cluster according to the length L of the target route and the maximum endurance mileage Lys of the artificial intelligence robot. The target cluster includes the inspection points for the artificial intelligence robot to conduct inspections; Generation module: Remove the inspection points in the target cluster, repeat the steps in the initial module and the analysis module, determine all the target clusters, and determine the inspection plan according to the target clusters. The inspection plan includes the number of artificial intelligence robots and the inspection paths; When the length L of the target route in the target cluster < 0.5Lys, perform the following steps: Determine the target cluster C with the length L of the target route < 0.5Lys, determine the distance between the center of the target cluster C and the centers of the remaining target clusters, and determine the shortest distance Dmin zx The corresponding target cluster Cdy; Take the target cluster C and the target cluster Cdy as the target cluster CCdy, determine the reference point K on the target route MB in the target cluster CCdy, and take all the monitoring points between the end points of the target route MB and the reference point K as a target cluster.

Images